DOCUMENTOS DE TRABAJO

Institutional Quality and Sovereign Flows

David Moreno

N° 816 Abril 2018BANCO CENTRAL DE CHILE

BANCO CENTRAL DE CHILE

CENTRAL BANK OF CHILE

La serie Documentos de Trabajo es una publicación del Banco Central de Chile que divulga los trabajos de investigación económica realizados por profesionales de esta institución o encargados por ella a terceros. El objetivo de la serie es aportar al debate temas relevantes y presentar nuevos enfoques en el análisis de los mismos. La difusión de los Documentos de Trabajo sólo intenta facilitar el intercambio de ideas y dar a conocer investigaciones, con carácter preliminar, para su discusión y comentarios.

La publicación de los Documentos de Trabajo no está sujeta a la aprobación previa de los miembros del Consejo del Banco Central de Chile. Tanto el contenido de los Documentos de Trabajo como también los análisis y conclusiones que de ellos se deriven, son de exclusiva responsabilidad de su o sus autores y no reflejan necesariamente la opinión del Banco Central de Chile o de sus Consejeros.

The Working Papers series of the Central Bank of Chile disseminates economic research conducted by Central Bank staff or third parties under the sponsorship of the Bank. The purpose of the series is to contribute to the discussion of relevant issues and develop new analytical or empirical approaches in their analyses. The only aim of the Working Papers is to disseminate preliminary research for its discussion and comments.

Publication of Working Papers is not subject to previous approval by the members of the Board of the Central Bank. The views and conclusions presented in the papers are exclusively those of the author(s) and do not necessarily reflect the position of the Central Bank of Chile or of the Board members.

Documentos de Trabajo del Banco Central de ChileWorking Papers of the Central Bank of Chile

Agustinas 1180, Santiago, ChileTeléfono: (56-2) 3882475; Fax: (56-2) 3882231

Documento de Trabajo

N° 816

Working Paper

N° 816

Institutional Quality and Sovereign Flows

David Moreno

Banco Central de Chile

Abstract

This paper evidences the relevance of mercantilism and sudden stops in emerging’market economies as a

joint explanation for positive net public foreign assets, while political-economy frictions account for the

varying degrees that asset accumulation is achieved across economies with similar characteristics. An

increase of 50 percent in these frictions implies a reduction of almost 20 percentage points of GDP in net

foreign public assets, in the context of economies with growth externalities, therefore a motive for

mercantilism exists. In its absence, on the other hand, the sovereigns accumulate net foreign liabilities,

which makes their access to international markets more infrequent. In a model without mercantilist

motives, political-economy frictions explain less of the differences in net foreign positions (5 percentage

points of GDP), but can explain great differences in sovereign risk: a reduction of 70 percent in political-

economy frictions can reduce the sovereign spread by 800 basis points.

Resumen

Este artículo documenta la relevancia del mercantilismo y las detenciones súbitas en las economías de

mercados emergentes como una explicación conjunta de posiciones internacionales netas positivas del

gobierno, mientras que las fricciones de economía política explican la variedad de grados de tal

acumulación de activos entre economías con similares características. Un aumento de 50 por ciento de

tales fricciones implica una reducción de casi 20 puntos porcentuales de PIB en activos internacionales

netos del sector público, en un contexto donde hay externalidades de crecimiento y motivos mercantilistas.

En ausencia de estos últimos, el soberano acumula, en cambio, pasivos netos internacionales, haciendo

que su acceso a mercados internacionales sea menos frecuente. En un modelo sin motivos mercantilista,

las fricciones de economía política explican menos de las diferencias de las posiciones netas (5 puntos de

PIB), pero pueden explicar grandes diferencias en riesgo soberano: una reducción de 70 por ciento de las

fricciones de economía política pueden reducir el spread soberano en 800 puntos base.

This paper corresponds to Chapter 1 of my PhD dissertation at University of Maryland, with some minor updates. I am grateful

for useful discussions with Sebnem Kalemli-Özcan, Anton Korinek, Felipe Saffie, Ethan Kaplan, Nils Gornemann, John Shea,

Pablo Cuba-Borda, Nuno Limão, Bora˘gan Aruoba, Luminita Stevens, John Haltiwanger, Rodrigo Heresi, Veronika Penciakova,

and participants of seminars at University of Maryland and Banco Central de Chile for their very useful comments. The views

expressed in this paper are exclusively those of the authors and do not necessarily reflect the position of the Banco Central de

Chile or its Board members. Any errors or omissions are the responsibility of the author.

Contact: dmoreno@bcentral.cl. Address: Agustinas 1180, Santiago 8340454, Región Metropolitana, Chile

1 Introduction

Global imbalances are an important feature of the international economy: high-growth

emerging-market economies export capital to advanced economies.1

These capital inflows are rooted in sovereign-to-sovereign flows, mostly international re-

serve accumulation and foreign public debt consolidation (Alfaro et al., 2014; Gourinchas and

Jeanne, 2013; Krishnamurthy and Vissing-Jørgensen, 2007). The literature has dealt with three

main reasons for this phenomenon: precautionary savings (Durdu et al., 2009; Jeanne and Ran-

cière, 2011), mercantilist motives (Benigno and Fornaro, 2012; Korinek and Servén, 2010), and

the assurance by the government against expropriation risk to foreign investors (Aguiar and

Amador, 2011; Aguiar et al., 2009). However, none of these three explanations alone can ex-

plain the extent and diversity of this phenomenon, and focusing in one mechanism at the time

does not allow the assessment of their relative importance.2

In this respect, this paper shows the great importance of mercantilist motives and sudden

stops in emerging market economies in explaining the existence of net positive international

public assets, while political economy frictions can account for the varying degrees that asset

accumulation is achieved across economies with similar characteristics. An increase of 50 per-

cent in these frictions implies a reduction of almost 20 percentage points of GDP in net foreign

public assets, in the context of economies with growth externalities, so there is a motive to en-

gage in mercantilism. On the other hand, if mercantilism is not taken into account, sovereigns

accumulate net foreign liabilities, which dampens their access to international markets. In a

model without mercantilism, political-economy frictions have less power to explain differences

in net foreign positions (5 percentage points of GDP), but can explain differences in sovereign

risk: a reduction of 70 percent in political-economy frictions can reduce the sovereign spread

by 800 basis points.

As far as I know, this is the first paper providing a plausible framework for a small open

1 For a specific definition and discussion on global imbalances, see Gourinchas and Rey (2014).2 More recently, Bianchi et al. (2012) have proposed a fourth motive, rollover risk, which will not be dealt with

in the present work.

1

economy where the government builds a positive net foreign asset position as an optimal

policy, while stressing the important difference that political-economy frictions and mercant-

ilist strategies make regarding net foreign-asset positions of governments in emerging-market

and developing economies.3

The present work builds a two-sector small-open-economy model featuring political-economy

frictions, sovereign default, exogenous sudden stops affecting firms’ working capital, and learning-

by-trade externalities. Political-economy frictions consist of time-inconsistent preferences ori-

ginating in the differences in consumption valuation of incumbent versus opposition parties,

which increases the risk that the government expropriates foreign assets in the country. For

this paper, instead of modelling the nationalisation of foreign capital, a sovereign default is

introduced, as default has been far more common during this era of globalisation than nation-

alisation, as shown by Tomz and Wright (2010). Sovereign default risk can be interpreted also

as expropriation risk, since the government may appropriate future debt payments belonging

to their creditors.

In this setting, firms in the tradable sector face occasional exogenous sudden stops, which

affect their working capital. In practical terms, they face financing shortages when purchasing

imported goods to be used in their production processes. When this happens, the government

may step in and provide credit, but in an inefficient and limited way, since it is usually subject to

sudden stops as much as their firms (Bianchi et al., 2012). This provides a strong precautionary

savings motive.

The presence of externalities in the tradable sector is due to the fact that the aggregate stock

of knowledge utilised to produce a modern tradable good has a positive covariance with the

amount of imported inputs utilised. However no firm can exploit this covariance privately. The

government cannot provide direct subsidies to the tradable sector to exploit these externalities,

3 Most papers yield a positive stock of foreign reserves (Benigno and Fornaro, 2012; Jeanne and Rancière, 2011)but without considering the stock of public debt; Benigno and Fornaro (2012) manages to explain jointly apositive net foreign asset position of the overall economy, but is silent about foreign public debt. Bianchi et al.(2012) manages to achieve a joint explanation of the existence of public debt and reserve assets, although thenet balance is still negative. Aguiar and Amador (2011) and Aguiar et al. (2009) explain why economies reducetheir foreign liabilities in order to reassure investors lower expropriation risk and promote growth, but the netforeign position of the government is still negative.

2

since such subsidies are forbidden by international trade law. The only instrument she has to

exploit this externality is the accumulation of net foreign assets. This motive of net foreign asset

accumulation by the government is called “mercantilism” and is still subject to debate in policy

and academic research, both on the empirical and theoretical front.

The next section features a literature review, which is followed by a discussion of the empir-

ical facts, the quantitative model and its properties, and the final results and conclusions.

2 Literature

The closest strand of literature to this chapter is that on small open economies with political-

economy frictions. The political-institutional setting in this paper comes directly from Aguiar

and Amador (2011). The authors propose a theory of political-economy frictions causing gov-

ernments to vary their net foreign assets to reduce the expropriation risk. This expropriation

consists of nationalising the capital brought in by foreign investors, enduring autarky forever

as a punishment. Hence, economies with better institutions (less political-economy frictions)

secure foreign investors’ property rights by increasing public foreign savings. This also pro-

motes faster convergence to the steady state. However, my paper is different in that it models

sovereign default: it considers the fact that expropriation of foreign creditors to the sovereign

are far more frequent than nationalisations of foreign physical capital; it also considers the

presence of sudden stops that arrive from abroad and reduces access to finance by the overall

economy, and adds a growth trend which is affected by the externalities present in the tradable

sectors.

Other papers focus on the government’s gross positions. As for gross debt, Amador (2003,

2012) and Aguiar and Amador (2014) assume that politicians demand debt ex-post due to

their inability to save stemming from political-economy frictions similar to those in Aguiar

and Amador (2011), so the desire to borrow again in the future enforces repayment today.

D’Erasmo (2008), using a different political structure, shows that a benevolent government

transiting between two states of patience replicates the observed default frequency and the

ratio of gross sovereign debt to gross national product at the moment of default.

3

For foreign gross sovereign assets (for which reserve accumulation is the main compon-

ent), literature linking reserves and institutions is rather scant. Aizenman and Marion (2004)

show both empirically and theoretically that corruption and short-term incentives reduce the

demand for foreign reserves, which is qualitatively consistent with papers addressing gross

foreign sovereign debt.

Our paper leaves as a future extension to consider gross positions of the government, both

for ease of modelling and exposition, an the impact on gross and net asset positions with respect

to measures of institutional quality are the same empirically, as expected.

Another literature worth mentioning is that studying institutional quality and fiscal counter-

cyclicity. Governments from countries with strong, savvy institutions tend to lower fiscal

spending during booms and increase it during busts, as shown in Frankel et al. (2013). Fiscal

counter-cyclicity depends on the level and changes in the measures of institutional quality.

Also, there is a literature linking institutional quality and private capital flows: Gourio et al.

(2015) shows that the VIX forecasts political risk, and that when it increases, capital inflows de-

crease and outflows increase. The expropriation risk in their model is a stochastic tax on capital

inflows, i.e. a nationalisation of foreign capital. Despite the fact that expropriation risk is not

endogenous, their work highlights the mechanisms by which institutional quality influences

private capital inflows, as found previously by Alfaro et al. (2008) and Papaioannou (2009).

Also, there is a vast literature on foreign reserve accumulation that addresses the basic eco-

nomic motives for governments to accumulate assets. Precautionary savings is one important

motive. International reserves provide a defence against sudden stops under increasing finan-

cial globalisation (Durdu et al., 2009). Other aspects are explored in Aizenman and Lee (2007),

Choi et al. (2007), Alfaro and Kanczuk (2009), and Caballero and Panageas (2008). Despite its

importance, the precautionary motive in isolation fails to account for a positive net foreign asset

position, which has been very common during the last decades, as governments increased their

net reserve hoarding and decreased their international sovereign debt positions. Explaining a

positive net foreign asset position is an important achievement of my paper.

4

Another important motive for reserve accumulation is mercantilism.4Some economies do

not let their currencies appreciate in order to promote economic growth through exports (Dooley

et al., 2004; Eichengreen, 2004). Accumulating reserves, hence, is a result of preventing real

appreciation, and this is how reserve accumulation can be related to output growth (Rodrik,

2008).5 Models of mercantilist reserve accumulation rely on endogenous growth with extern-

alities present in the tradable sector. To exploit these externalities, the central planner hoards

reserves as a second best mechanism to direct subsidies (Korinek and Servén, 2010). As a by-

product, these reserves also provide foreign working-capital flow during sudden stops (Be-

nigno and Fornaro, 2012).6

These models describe the mechanisms by which mercantilism works, but explaining why

successful cases of mercantilism are so few requires looking at political and institutional factors.

There are well-studied cases of failed export-led growth strategies, such as Latin America

between 1870 and 1930 (Catão, 1992; Cortés-Conde, 1992; Gómez-Galvarriato and Williamson,

2009), where political issues were connected with the failure of these policies.7,8 In this paper, I

show that political-economy frictions can account for differences in the amount of foreign sav-

ings across government. Also, political economy frictions can explain significant differences in

the output per capita in the long run.

Other motives for hoarding reserves which I do not address are their role in enhancing

domestic policies by allowing exchange-rate manipulation (Alfaro and Kanczuk, 2013), and the

4 Mercantilism refers to the prevalent economic doctrine in Europe during the XVII-XIX centuries, where coun-tries accumulated monetary reserves (mainly gold) by running trade balance surpluses of finished goods. Fora reference, see Ekelund and Hébert (2013).

5 This view, however, has been challenged by Reinhart et al. (2016), as reserve accumulation seems to crowd outprivate investment in East-Asian economies, excluding China and India (which coincidentally feature strictcapital controls.) Woodford (2009) argues, on the other hand, that undervaluation is just a consequence ofexcess savings and strict capital controls, instead of means to promote growth by exporting.

6 An important matter is why reserves are held despite yielding lower returns compared to other instruments.One reason is its imperfect substitutability with private foreign debt (Benigno and Fornaro, 2012). Intermittentaccess to international capital markets leaves reserves as the sole backstop to tradable-goods producing firms.Under perfect substitutability, it is optimal to reduce debt holdings and hold no foreign reserves (Alfaro andKanczuk, 2009).

7 After the Great Depression, most countries in Latin America embarked in Import-Substitution strategies, fol-lowing the “Prebisch Doctrine” (Prebisch and Martínez-Cabañas, 1949). For a contrast with Emerging Asia, seeBaer (1984).

8 For a description of the prevailing international monetary system prevailing during 1870-1930 see Eichengreen(1992) and Obstfeld and Taylor (2003).

5

prevention of rollover risk in long-term sovereign foreign debt (Bianchi et al., 2012). The former

requires a monetary model, which is beyond the scope of this paper, while the latter requires

modelling gross positions, which is left to future extensions.

In a broader sense, the current paper relates to the global imbalances literature, which has

been prolific and well-cited. The first theories emphasised the presence of a global savings glut

(Bernanke, 2005) –although with no regard to which sectors actually engaged in saving– while

connecting this to the conundrum of low long-term interest rates in the United States (Green-

span, 2005). On the other hand, Dooley et al. (2004) and Eichengreen (2004) explain global

imbalances instead using mercantilism, by comparing the current global imbalances (1996 to

present) to those of the Bretton Woods era (1946-1973). Countries pursuing export-led growth

strategies –such as Germany and Japan at that time– resisted depreciation of their currencies,

just as Emerging Asia is thought as doing recently.9

Another view is financial sector development, which is low in economies less able to di-

versify away idiosyncratic risk (Angeletos and Panousi, 2011; Buera and Shin, 2009; Mendoza

et al., 2009), which explains why firms in less financially developed countries export capital

to firms in more financially developed economies. Although some supporting evidence exists

for advanced economies (Mendoza et al., 2009), for the rest of the world Alfaro et al. (2014)

and Gourinchas and Jeanne (2013) show that these savings were channelled abroad by the pub-

lic sector rather than the private sector. Moreover, Chinn et al. (2014) shows that government

budget deficits in advanced economies and public savings in emerging-market economies are

related under global imbalances. This is reserve accumulation in form of safe instruments such

as Treasuries, Gilts, BTFs, and Bunds. As shown in my paper, emerging-market countries with

better institutional will demand more foreign reserve assets. This demand increase the extent

to which global safe assets are scarce (Caballero et al., 2008), and hence strengthen global im-

balances.

Although less directly, this paper relates to work on the economic effect of institutions such

as Acemoglu et al. (2001), Acemoglu et al. (2014), Alesina and Dollar (2000), Jones and Olken

9 For a general review on possible causes of global imbalances, refer to Eichengreen and Park (2006) and Eichen-green (2006).

6

(2005), Jones and Olken (2009), Lindqvist and Östling (2010) and Azzimonti and Talbert (2014),

among other works.

The following section will show stylised facts on the relation between institutional quality

and foreign public assets in the data.

3 Stylised Facts

This section shows the relationship between institutional quality and public net and gross

foreign assets. Data come primarily from the World Development Indicators Database Archive.

Net foreign savings is measured as the difference between stock of international reserves (ex-

cluding gold) and the stock of Public and Publicly Guaranteed Debt. To extend data availability,

I use vintages of the World Development Indicators (WDI) and Global Development Finance

(GDF) issues dating back to 1989, as some countries are dropped from the dataset as soon as

they become high-income economies, according to the World Bank classification.

Data on national accounts, purchasing-power-parity (PPP) measures and foreign exchange

rates come from the Penn World Tables (PWT) version 8.1 (Feenstra et al., 2015). This new

version makes available data on national accounts at current and constant local-currency prices

in addition to the PPP-adjusted series.

Another important set of indicators for this topic are the capital account openness measures

from Chinn and Ito (2008), and the Political Risk Index and its subcomponents, from Interna-

tional Country Risk Guide (ICRG), available monthly since 1984.10 I will particularly use the

investment profile subcomponent as a measure of expropriation risk. The investment profile cat-

egory is an assessment of factors affecting the risk to investment that are not covered by other

political, economic, and financial risk components in the total ICRG index of Political Risk, and

it is composed of three subcategories: contract viability/expropriation, profits repatriation, and

payment delays. The first measures the risk of unilateral contract modification or cancellation

and, at worst, outright expropriation of foreign owned assets; the second measures to what ex-

10 Another set of indices widely used in the literature are those related to political structure and are much longerdated, for example, Polity IV, Freedom House, Keefer and Stasavage (2003), and Beck et al. (2001). However,they do not measure expropriation risk as directly as some subcomponents of ICRG.

7

tent can profits be transferred out of the host country (impediments include exchange controls,

excessive bureaucracy, a poor banking system, etc.); and the latter is the risk associated with

receiving and exporting payments from the country (impediments include poor liquidity, ex-

change controls, an inadequate banking system, etc.). Although the first subcomponent is the

most direct measure of expropriation risk, it is available only starting 2001. For this reason I use

the broader investment profile time series, dating back to 1984.

Ideally, we would like to make series as comparable as possible, both internationally and

intertemporally. Following Gourinchas and Jeanne (2013), I use an implicit trade deflator QTct

for country c and year t,11 from PWT 8.1, as a combination of the export and import deflators,12

weighted by their shares in GDP:

QTct =

Xct

Xct + MctQX

ct +Mct

Xct + MctQM

ct (1)

where X and M corresponds to the average-of-period shares in current PPP-adjusted exports

and imports, respectively. In previous versions, this trade deflator is not available, so the in-

vestment deflator is used to compare across countries; however, to compare across periods, the

investment deflator is adjusted further with the GDP deflator:

QTct = QT

ct ×CGDPct

RGDPct(2)

where CGDP (RGDP) is the GDP in current (chained) PPP-adjusted 2005 dollars. The adjus-

ted stocks of foreign assets and liabilities, which I denote as real stocks, are scaled by GDP in

chained PPP 2005 dollars. The results are reserves and PPG debt adjusted across period and

country, and scaled by real GDP.

In Figure 1 I depict the relation between net sovereign foreign assets and expropriation risk

(as measured by investment profile component of the ICRG Political Risk Index.) The relation is

11 They use version 7.1 of PWT, which does not report deflators for exports and imports, so the authors deflateseries using the price of investment; however, they recommend the use of trade deflators for capital flowswhenever possible.

12 These deflators are normalised such that the price level of GDP in the United States in 2005 is equal to 1.

8

significant and positive. When decomposing the net foreign public assets into foreign reserves

and PPG debt, as shown in Figure 2, it is clear PPG debt is negatively correlated with the index

of expropriation risk, while foreign reserves are positively correlated. All of these relations are

maintained when controlling for outliers.

It is important, however, to control for other determinants of the public international in-

vestment position. I will follow the empirical strategy of Gourinchas and Jeanne (2013). This

strategy follows from a neoclassical model that, although differing from the model I build in

this chapter, features also a small open economy.

Their right hand side variables consist of international capital flows, both private and pub-

lic. These are accumulated flows between 1980 and 2000, scaled by GDP. I will focus on the

public flows, which are composed of Public and Publicly Guaranteed (PPG) Debt and Interna-

tional Reserves, as explained in previous paragraphs. On the left hand side I will control for

productivity catch-up with the United States, the ratio of initial capital to GDP, and the ratio of

foreign public debt to GDP, the growth of the 15-64 year-old population, capital openness (as

measured by Chinn and Ito, 2008), and the expropriation risk measure from ICRG.13

The results are shown in Table 1. The results from columns 1 to 4, reproduce the original

results. The coefficients must be interpreted with the opposite sign, this is, an increase in pro-

ductivity catch-up will increase the outflows of the public flows (recall that in the Balance of

Payment Manual 5 an outflow is indicated with a negative sign), as does an increase in capital

account openness.

Columns 5 to 8 add the expropriation risk measure as independent regressor, where results

suggests that expropriation risk is a fundamental factor for public flows, affecting the public

foreign savings mainly through reserve accumulation. An increase of one standard deviation

in the measure of expropriation risk, the investment profile subcomponent of the ICRG polit-

ical risk index increases net foreign assets in 20 pp of GDP. In other words, an increase in the

investment profile subcomponent, which corresponds to a lower expropriation risk, implies a

13 These regressions use the original data, available from their publisher, except for the investment profile variablefrom ICRG. For a description on these variables, data, and countries included in the sample, refer to Gourinchasand Jeanne (2013).

9

greater accumulation of public net foreign assets of 20 pp of GDP.

Columns 9 to 12 show the results adding an interaction between the productivity catch-up

and expropriation risk, which is significant at the 10 percent level, while the result is only sig-

nificant at levels of 5 percent through foreign reserves. This indicates that lower expropriation

risk by the government magnifies the effect of higher productivity catch-up in the accumulation

of public net foreign assets.

The next section describes the model I use to account for the expropriation risk effects on

net foreign public savings, under sudden stops and default risk.

4 Model

I consider an infinite-horizon small open economy, where time is discrete and indexed by

t, populated by a continuum of mass 1 of households and by a large number of firms. The

firms are owned by the households. Some produce tradable goods, and the rest, nontradable

consumption goods. The government is run by one of several political parties, which is elected

at the beginning of each period t. A political-economy friction leads political parties to value

consumption more as an incumbent rather than as opposition, generating less-than-optimal

policies. The government is the only agent engaged in borrowing/saving in the international

capital markets with foreign investors. In every period the government can repay its debt or

default on it. The following subsections provide further detail on each aspect of the model.

4.1 Political Environment

The political environment of this economy follows Aguiar and Amador (2011). There is a set

I = 1, 2, 3, · · · , N + 1, where N + 1 is the number of parties. The government is controlled

by an incumbent party, chosen at the beginning of every period from the set I . This party may

lose and come back into power eventually. The key assumption is that the incumbent party

strictly prefers consumption occurring under its rule.

10

Assumption 1 (Political Economy Friction). A party enjoys a utility flow θu (Ct) when in power and

a utility flow u (Ct) when out of power, where C is the per-capita consumption basket by the domestic

households and where θ > 1.

There are several possible interpretations for the parameter θ. In principle, this parameter

captures difference in intertemporal comparisons between the party in office and the opposi-

tion. One interpretation is disagreement regarding government expenditures. Another one is

corruption where the ruling party captures a disproportionate share of consumption per-capita.

The transfer of power is modelled as an exogenous Markov process. Denote p as the probab-

ility that the party in office retains power. If the party in office loses, each party in the opposition

has an equal probability of gaining power. Denote q as the probability of regaining power, i.e.

q ≡ (1− p) /N. In particular, (p− q) ∈[−N−1, 1

]represents the incumbent advantage on

elections. Denote pt,s as the probability in period t that the incumbent will be in office in period

s > t. A Markov political process can be represented as:

pt,s+1 = p× pt,s + q× (1− pt,s) (3)

Starting from pt,t = 1, this equation has the following solution:

pt,s = p + (1− p) (p− q)s−t (4)

where p = lims→∞ pt,s is the unconditional probability of taking office. For p < 1, p =

(N + 1)−1, while for p = 1, p = 1. As a consequence of the political process and the political-

economy friction, the utility of the incumbent can be written as:

Wt = Et

[∞

∑s=t

βs−t pt,sθu(Cs) +∞

∑s=t

βs−t (1− pt,s) u(Cs)

](5)

11

For ease of analysis, let us define the following ratios:

θ ≡ θ

pθ + (1− p)(6)

δ ≡ p− q =p− p1− p

(7)

The parameter θ is the ratio of the conditional valuation of consumption flow during incum-

bency over the unconditional valuation when in the opposition. The parameter δ will represent

the incumbency advantage, and can also be interpreted as the persistence of θ in the planning

horizon of the incumbent. Using these new parameters, Equation (5) can be rescaled using

equations (6) and (7), yielding the following representation of the incumbent preferences:

Wt ≡Wt

p(θ + N

) = Et

[∞

∑s=t

βs−t (θδs−t + 1− δs−t) u (Cs)

](8)

The differences between the incumbent’s and the opposition’s preferences stem from the

discounting processes. The intertemporal discount factor varies over periods, which is dif-

ferent from the case of a constant discount rate β. Between t and t + 1 the discount factor is

β (θδ + 1− δ) /θ and between t + 1 and t + 2 it’s β(θδ2 + 1− δ2) / (θδ + 1− δ). Only as t→ ∞

does the intertemporal discount rate converge to β. This is a feature of quasi-hyperbolic dis-

counting à la Laibson (1997), and the comparison is exact when δ = 0 and p = q = (1 + N)−1

(no incumbency advantage), in which case preferences are given by:

Wt = u (Ct) +1θ

Et

∞

∑s=t+1

βs−tu(Cs) (9)

where the discount factor between period t and t + 1 is β/θ, and afterwards β > β/θ. The

parameter δ > 0 makes the differences in discount rates persist over time, given θ > 1. Notice

as well that autocratic governments, where δ = 1, will yield the same results as the case where

θ = 1, so any differences in consumption/saving behaviour of their economies will come from

the degree of impatience they exhibit, represented by parameter β. In the following sections I

will describe households, firms, and further details on the public sector.

12

4.2 Households

The representative household derives utility from consumption Ct and supplies labour in-

elastically each period. The household’s lifetime expected utility is given by

Et

[∞

∑s=t

βs−t C1−σs

1− σ

](10)

In this expression, Et is the expectation operation conditional on information available at time

t, β < 1 is the subjective discount factor, σ > 0 is the coefficient of relative risk aversion and Ct

denotes a consumption basket. This consumption basket is a constant-elasticity-of-substitution

(CES) combination of tradable CTt and nontradable CN

t goods. The parameters are a, which is

the share of income destined to buy tradable goods, and ζ, which is the inverse of one minus

the elasticity of substitution between tradable and non tradable goods:

Ct =

[a(

CTt

)−ζ+ (1− a)

(CN

t

)−ζ]−1/ζ

(11)

Each period the household faces the following flow budget constraint:

CTt + PN

t CNt = Wt + ΠT

t + PNt ΠN

t + Tt (12)

The budget constraint is expressed in units of the tradable good. The left-hand side repres-

ents the household’s expenditure, where PNt represents the relative price of the nontradable

good in terms of the tradable good, so CTt + PN

t CNt is the household’s consumption expenditure

expressed in units of the tradable good. The right-hand side represents the income of the house-

hold. Wt denotes the household’s labour income. ΠTt and ΠN

t are the dividends the households

receives from firms operating in the tradable and in the nontradable sector, respectively. Tt rep-

resents the net tax/transfers to/from the government after its net asset decisions are made. For

simplicity, domestic households do not trade directly with any foreign investors. Each period

the representative household chooses CTt and CN

t to maximise expected utility (10) subject to

13

the budget constraint (12). The first order condition is:

(1− a

a

)(CT

tCN

t

)1+ζ

= PNt (13)

where PNt will be considered as a proxy for the real exchange rate. The consumer-price index is

given then by

Pt =

[a−ζ + (1− a)−ζ

(PN

t

)1+ζ] 1

1+ζ

(14)

which is the composite price-index of the consumption basket Ct.

4.3 Firms in the tradable sector

Firms in the tradable sector produce the final tradable good YTt using labour LT

t and impor-

ted intermediate goods Mt. The production function is Cobb-Douglas with labour share α, a

labour-augmenting productivity factor Γt, and a temporary technology shock zt.

YTt = zt(ΓtLT

t )αM1−α

t (15)

The labour-augmenting productivity factor Γt is the knowledge used to produce the tradable

good, which is public and non-rival, and the transitory technology shock zt has its logarithm

autocorrelated of first order with persistence ρ and white noise innovations εt, whose mean is

0 and its standard deviation, σε.

log zt = (1− ρ) log µz + ρ log zt−1 + εt (16)

Eεt = 0, Eε2t = σ2

ε

A constant fraction φ of the purchases of imported intermediate goods must be financed with

foreign intra-period loans every period, up to a stochastic borrowing limit κt. In addition, the

government can provide public loans Dt in case the constraint binds, and foreign financing is

14

not sufficient to purchase the intermediate imported inputs:

φPMt Mt ≤ κtΓt + Dt (17)

This borrowing limit κt can take two values: κL and κH > κL; and follows a Markovian discrete

process with transition probability F(κt|κt−1). The higher value κH is enough to ensure that the

borrowing constraint is never binding, while κL makes the borrowing constraint bind under

certain states. Whenever the latter arises, the government will provide public loans, subject to

its availability of funds. The profit function for the entrepreneurs is given then by:

ΠTt = zt

(ΓtLT

t

)αM1−α

t − PMt Mt −WtLT

t + µt

(φPM

t Mt − κtΓt − Dt

)(18)

where µt stands for the multiplier of the borrowing constraint. The first order conditions for

the tradable firms are:

ztα (Γt)α(

LTt

)α−1 (MT

t

)1−α= Wt (19)

zt(1− α)(

ΓtLTt

)α (MT

t

)−α= PM

t (1 + φµt) (20)

µt

(φPM

t Mt − κtΓt − Dt

)= 0 (21)

It is important to highlight that even if the government had the resources to step in, the

multiplier µt would still be positive unless it can restore the full first-best choice of the firm.

4.4 Knowledge accumulation process

The stock of knowledge available to firms in the tradable sector evolves according to the

following process:

Γt+1 = υΓt + Mξt Γ1−ξ

t (22)

where υ ≥ 0 and 0 ≥ ξ ≥ 1. This formulation captures the idea that imports of foreign

capital goods represent an important transmission channel through which discoveries made in

15

developed economies spill over to developing countries. As mentioned before, knowledge is

assumed non-rival and non-excludable. This, along with the large number of firms assumed

in the tradable sector, implies that firms do not internalise the impact of their actions on the

evolution of the economy’s stock of knowledge. We can rewrite the equation of the evolution

of knowledge as follows:

gt+1 ≡Γt+1

Γt= υ +

(Mt

Γt

)1−ξ

(23)

where gt+1 is the growth rate of the knowledge stock.

4.5 Firms in the nontradable sector

The nontradable sector represents a traditional sector that does not engage in international

trade. Its output is produced using labour according to the function:

YNt = Γt(LN

t )γ (24)

where Γt is the growth of the stock of knowledge in the economy, and 0 ≤ γ ≤ 1 is the share of

labour in profits. The first order condition for labour is:

γPNt Γt

(LN

t

)γ−1= Wt (25)

In addition, the labour market must clear LTt + LN

t = 1, as must the market of non-tradable

goods CNt = YN

t .

4.6 Private Sector Equilibrium

Definition 1 (Private Sector Equilibrium). A Private Sector Equilibrium is characterised by a set of

allocations CTt , CN

t , LTt , LN

t , Mt∞t=0, and prices Wt, PN

t ∞t=0 such that, taking as given government

policies Tt, Dt∞t=0, knowledge process Γt∞

t=0, and stochastic processes κt, zt∞t=0:

(i) Households satisfy Equation (12) and Equation (13), taking as given prices, profits from tradable

firms ΠTt , profits from nontradable firms ΠN

t , and government lump sum net transfers Tt.

16

(ii) Firms satisfy Equation (19),Equation (20), Equation (21) and Equation (25) and satisfy labour and

non-tradable goods market clearing, taking as given prices, the knowledge process, and government

policies.

For a solution of the private sector equilibrium, see Section A.

4.7 The Government Budget Constraint

The sovereign can issue one-period, non-contingent discount bonds, so contingent claims

markets are incomplete. Alternatively, it can buy other one-period, non-contingent discount

bonds in the foreign market, which are risk-free. The face value of these bonds specifies the

amount to be repaid/received in the next period, Bt+1. The government borrows if Bt+1 < 0

and saves if Bt+1 > 0. The set of the net government savings is thus B ⊂ R.

The lower bound b is usually set to be higher than −GDPT/r∗, an annuity of tradable value

added, which is the largest debt that the country could repay under full commitment. Altern-

atively, the upper bound b can be set lower than GDPT/r∗, which is the largest savings that the

country can accumulate. Hence B =[b, b], and 0 ∈ B. The price of these bonds is qB

t which is a

function qB (Bt+1, Γt, zt, κt) set by foreign investors.

In addition, the government can provide tradable-goods producing firms with intraperiod

working capital loans Dt when they are under distress. However, when providing these loans,

the government incurs on efficiency loss, which amounts to ψ/ (1− ψ) Dt (Gertler and Karadi,

2011). Therefore, the government budget constraint is given by:

Bt = Tt + qBt Bt+1 +

ψ

1− ψDt (26)

where Tt is the amount of net transfers to the households. The loans to the private sector cannot

exceed the net savings the government has at the moment, so:

Dt ∈ [0, (1− ψ) Bt]

The public loans to the private sector are thus the minimum between the amount that tradable-

17

goods producing firms need to finance the unconstrained imported inputs purchases, and the

total amount of resources the government can provide to this sector. Hence, the rule is given

by the following equation:

Dt = max

min

φPMt Mu

t − κtΓt, (1− ψ) Bt

, 0

(27)

where Mut is the unconstrained level of imported inputs used by the firms in tradable sector, φ

is the fraction of imported inputs financed with short-term foreign loans, and κ is a credit shock

to the foreign borrowing limit the tradable firms face, scaled by the technology long-term trend

Γt. An additional assumption, common in the literature (see Bianchi et al., 2012) is that the

government cannot borrow in the international markets during a sudden stop.14

4.8 Default

The sovereign cannot commit to repay its debt. As in Eaton and Gersovitz (1981), when

the country defaults it does not repay at date t and is excluded from the world credit markets

starting the same period. The country may re-enter into the international capital market with

an exogenous probability η, starting with a fresh record and zero debt.

The government chooses a saving/borrowing policy and whether to default or not, taking

the private sector decisions as given, according to the following rule:

Wot = max

Wc

t , Wdt

(28)

where Wct is the expected utility for the incumbent at t of not defaulting, and Wd

t is the expected

utility of defaulting; Wot is the maximum between the expected utilities of defaulting and not

14 A sudden stop in this paper is defined as a situation where government resources are not enough to restorefirst-best levels of unconstrained imported inputs. In other words, when κt = κL, the government may have tointervene but will not have enough resources.

18

defaulting. The government will default whenever Wct ≤Wd

t . The expected utilities are:

Wct,t = θu (Ct) + βEt

[Wo

t,t+1]

(29)

Wdt,t = θu

(Cd

t

)+ βEt

[ηWo

t,t+1 + (1− η)Wdt,t+1

](30)

Notice that Wot,t+1 corresponds to the value function of the incumbent in period t + 1 but from

the perspective of period t. Since the government does not commit and has naïve hyperbolic

discounting (as explained in section 4.1), it is very likely that Wot,t+1 6= Wo

t+1,t+1. We can write

down the value functions for any horizon h ≥ 0 as follows:

Wot,t+h = max

Wc

t,t+h, Wdt,t+h

(31)

Wct,t+h =

(θδh + 1− δh

)u (Ct,t+h) + βEt+h

[Wo

t,t+h+1]

(32)

Wdt,t+h =

(θδh + 1− δh

)u (Ct,t+h) + βEt+h

[ηWo

t,t+h+1 + (1− η)Wdt,t+h+1

](33)

Hence the probability of default depends on the likelihood that Wct,t+h < Wd

t,t+h. Denote this

probability Υt,t+h ≡ Pr[Wc

t,t+h < Wdt,t+h

]. This probability will be used by foreign investors to

assess the value of the government portfolio.

4.9 Foreign investors

International creditors are risk-neutral and have complete information. They invest in one-

period sovereign bonds and in within-period private working capital loans. Foreign lenders be-

have competitively and face an opportunity cost of funds equal to r∗. Competition implies zero

expected profits at equilibrium and full arbitrage between the sovereign debt and the world’s

risk-free asset. Hence, the price of the sovereign net international investment position is given

by:

qBt =

1/(1 + r∗) if Bt+1 ≥ 0

(1− Υt+1)/(1 + r∗) if Bt+1 < 0(34)

19

where Υt+1 is the probability of sovereign default in the next period. This result assumes that in

sovereign default, all assets and liabilities are seized, while whenever foreign assets are greater

than foreign liabilities, the sovereign does not default, as it can repay debt using those assets.

This provides an important simplification, although the confiscation of foreign assets seldom

occurs (Wright, forthcoming).

4.10 A Recursive Formulation

For a recursive formulation, we first denote s = κ, z as the vector of exogenous state

variables, while B and Γ are the endogenous state variables. At any given horizon h ∈ N0, the

value functions for continuation Wch, default Wd

h , and for the default option Woh are given below:

Wch(B, Γ, s) = max

B′Ωhu (Ch) + βEhWo

h+1(B′, Γ′, s′) (35)

Wdh (Γ, s) = Ωhu

(Cd

h

)+ βEh

[ηWo

h+1(0, Γ′, s′) + (1− η)Wdh+1(Γ

′, s′)]

(36)

Woh(B, Γ, s) = max

Wc

h(B, Γ, s), Wdh (Γ, s)

(37)

where we define Ωh ≡ δhθ + 1− δh. For a tractable recursive representation it is important to

consider that as t goes to infinity, the discount rate between consecutive periods, βΩh+1/Ωh,

tends to β. By using the definition of limits, it is possible to find a horizon τ such that for

any h > τ, the discount rate between any two consecutive periods is always β as in a classical

sovereign default problem.

Proposition 1. For a sufficiently small ε, θ > 1, and δ ∈ (0, 1), there exists τ such that for h > τ:

∣∣∣∣Ωh+1

Ωh− 1∣∣∣∣ < ε

and hence limh→∞

Ωh+1/Ωh = 1.

Proof. Define τ = logδ ε− logδ [(θ − 1) (1− δ− ε)]. Hence for h > τ the intertemporal discount

rate in subsequent periods is close enough to β.

20

Note that also τ is such that Wh = V for any h ≥ τ, where V is the value function of

households. Note that if δ = 0, then τ = 0, and W1 = V; if δ = 1 and θ ≥ 1, Wh = V, ∀h ∈N0.

As an illustration, in fig.3 the discount rate βhΩh(δ, θ) is depicted for various values of δ,

given parameters β = 0.95 and θ = 1.5. Note that the discount rate for the first period converges

in h = 1 if δ = 0, to the value that a benevolent sovereign would discount. When δ = 0.25,

Ωh converges after h = 10 for ε = 10−6, and after h = 41 for δ = 0.75. The importance of

this discount function discussion is that it allows us to solve the problem by using backward

induction. We solve for a benevolent sovereign for h ≥ τ and then solving backwards using

the corresponding factor Ωh. Using this result, for horizons h > τ we can solve a classical

sovereign-default recursive problem:

Vc(B, Γ, s) = maxB′

u (C) + βEVo(B′, Γ′, s′) (38)

Vd(s) = u(

Cdef)+ βE

[ηVo(0, Γ′, s′) + (1− η)Vd(Γ′, s′)

](39)

Vo(B, Γ, s) = max

Vc(B, Γ, s), Vd(Γ, s)

(40)

The definitions of the default set and the probability of default are standard from Eaton-Gersovitz

models (Arellano, 2008). For a debt position B < 0, default is optimal for the set of realisations

of s for which Vd(B, Γ, s) is at least as high as Vc(B, Γ, s):

D (B) =

s : Vc(B, Γ, s) ≤ Vd(B, Γ, s)

(41)

The probability of default at t + 1 perceived at t, Υ(B′, Γ, s), can be deduced from default set

and transition probability function F of shocks to productivity z and credit κ, as follows:

Υ(B′, Γ, s) =∫D(B′)

dF(s′, s) (42)

The transition probability for technology shocks z is continuous G(z′, z), given by Equation (16),

21

while that for credit shocks κ is discrete, and given by H(κ′, κ):

Υ(B′, Γ, s) = ∑κ′∈D(B′)

∫z′∈D(B′)

h(κ′, κ)g(z′, z)dz′ (43)

With this probability in mind, the price function of the sovereign portfolio is calculated as:

qB (B′, Γ, s)=

1− Υ (B′, Γ, s)1 + r∗

(44)

where Υ = 0 for all B′ ≥ 0.

After having solved the “terminal” problem, we proceed to iterate backwards until we ar-

rive to horizon h = 0. An important assumption is that foreign investors know whether the

government behaves hyperbolically or not, so they will calculate the default rule as:

D0 (B) =

s : Wc0(B, Γ, s) ≤Wd

0 (B, Γ, s)

(45)

Υ0(B′, Γ, s) = ∑κ′∈D0(B′)

∫z′∈D0(B′)

h(κ′, κ)g(z′, z)dz′ (46)

qB (B′, Γ, s)=

1− Υ0 (B′, Γ, s)1 + r∗

(47)

which is possible, because bonds are issued at one-period maturity.

4.11 Detrended Form

To solve the recursive formulation numerically, it is important to remove the productivity

trend Γ, as to reduce the number of states in the economy. This is a feasible procedure, given

that the value functions are homogeneous of degree 1− σ in Γ, and that the price function of

the sovereign portfolio is homogeneous of degree zero in Γ and B′. This derives from the utility

function specification and the fact that the budget and borrowing constraints are homogeneous

22

of degree one in Γ. The detrended form is given hence by:

vc(b, s) = maxb′

u (c) + β(

g′)1−σ

Evo(b′, s′) (48)

vd(s) = u(

cdef)+ β

(g′)1−σ

E[ηvo(0, s′) + (1− η)vd(s′)

](49)

vo(b, s) = max

vc(b, s), vd(s)

(50)

and g′ ≡ Γ′/Γ = υ + mξ is the future growth rate of the economy, which is known in current

period, although not internalised by the private sector. It is important to add that the constraints

for the private sector equilibrium are also scalable:

cT + PNcN = zm1−αLαT + PN(1− LT)

γ − PMm + b− qB(b′, s)b′g′ − ψ

1− ψd (51)

φPMm ≤ κ + d (52)

d = max

min

φPMmu − κ, (1− ψ)b

, 0

(53)

The resource constraint of the economy is given by equation (51), while the borrowing con-

straint is given by (52).

The solution of the model is equivalent to a constrained-centralised solution where the cent-

ral planner (the sovereign) takes the growth rate of the economy g′ as given, yielding the same

first order conditions of the private sector equilibrium. Once the terminal value functions and

policy rules are solved, we iterate backwards and find the solutions for shorter horizons, start-

ing in period h = bτc until h = 0:

wch(b, s) = max

b′Ωhu (ch (b, s)) + β

(g′)1−σ Ehwo

h+1(b′, s′) (54)

wdh(s) = Ωhu

(cd

h (s))+ β

(g′)1−σ Eh

[ηwo

h+1(0, s′) + (1− η)wdh+1(s

′)]

(55)

woh(b, s) = max

wc

h(b, s), wdh(s)

(56)

23

where the private sector constraints are:

cTh + PN

h cNh = zm1−α

h (LTh )

α + PN(1− LTh )

γ − PMmh + b− qBh (b′, s)b′hg′−

− ψ

1− ψdh (57)

φPMmh ≤ κ + dh (58)

dh = max

min

φPMmuh − κ, (1− ψ)b

, 0

(59)

5 Economy-Wide Equilibrium

Definition 2 (Recursive Equilibrium). A Recursive Equilibrium is characterised by a set of value

functions woh(b, s), wc

h(b, s), wdh(s)

τh=0 and vo(b, s), vc(b, s), and vd(s); a default rule Υh(b, s)τ

h=0 , Υ(b, s),

a sovereign portfolio rule

b′h(b, s)τ

h=0 , b′(b, s), a sovereign portfolio price function

qBh (b, s)

τ

h=0 , qB(b, s),

a credit policy dh(b, s)τh=0, d(b, s), and a transfer policy th(b, s)τ

h=0 , t(b, s), such that:

(i) Policy rules

b′h(b, s)τ

h=0 , b′(b, s) and dh(b, s)τh=0 , d(b, s) solve the problem in (48)-(59),

given the price function

qBh (b, s)

τ

h=0 , qB(b, s). That is the government’s default and borrow-

ing decisions are optimal given the interest rates on sovereign debt.

(ii) Private consumption and factor allocations are feasible and consistent with the equilibrium private

market defined in Section 4.6.

(iii) The transfer policies th(b, s)τh=0, dh(b, s)τ

h=0, t(b, s), and d(b, s) satisfy the budget con-

straints of the government.

(iv) Given default regions D(b, s)τh=0 and D(b, s), and probabilities of default Υ(g, s)τ

h=0, and

Υ(g, s), the bond price functions

qBh (b′, s)τ

h=0, and qB(b′, s) satisfy the arbitrage condition of

investors in equation (34).

A solution to the recursive equilibrium includes solutions for sectoral factor allocations and

production during normal periods and default, and sudden stops as well. Solutions for equi-

librium wages, profits, and the price of domestic inputs follow then from the firms’ optimality

conditions and the definition of profits described earlier.

24

6 Parametrisation

The parameters chosen are shown in Table 2. Parameters σ, β, and r∗ have standard RBC

values of 2, 0.9, and 4 percent (annual), respectively; for consumption, the parameters a and ζ

are taken from Schmitt-Grohé and Uribe (2016), although a has the same value as in Benigno

and Fornaro (2012). A value of ζ = 0.2 corresponds to a value of the constant elasticity of

substitution of 5/6.

For the trend process in the tradable sector, I set υ = 0.17 so as to match the growth rate

in a non-stochastic steady state to 3.4 percent, which is the average growth of emerging and

developing economies. The elasticity of growth to imported inputs is 0.1 and is slightly lower

than in Benigno and Fornaro (2012), although allows for the benchmark economy to have 30

percent of GDP in public net foreign assets. For the transitory process ρ = 0.96 and σε =

1.21%, and µz = 0.49 is set so total GDP in the non-stochastic steady state is equal to 1. These

parameters were calibrated for the Chilean economy, although they are very close to those in

Mendoza and Yue (2012) for the Argentinian economy.

For the firms, the share of labour in tradable and nontradable sectors, namely α and γ, are

both set to 0.6, which is a standard value. The price of imported inputs is normalised to 1.

The share of imported inputs that is financed with imported inputs, φ, is set to 0.7 (Mendoza

and Yue, 2012), while the social loss parameter ψ when the government provides credit, is 0.5

(Benigno and Fornaro, 2012).

The transition probabilities for the borrowing limit κ are such that the economy enters into

a bad shock every 10 years, and stays there for 2 years (Benigno and Fornaro, 2012). The low

value κL is set to 0.12, which yields a trend growth rate of -2.5 percent during its occurrence.

If the sovereign defaults, the probability of reentry is 1/3, which means that the economy

stays an average of 3 years without accessing the international capital markets (Bianchi et al.,

2012). The political-economy friction parameter θ is set to different values, ranging from 1 to

1.5, while I consider that the incumbency advantage is 0, as in Aguiar and Amador (2011), for

ease of calculation.

25

7 Computation strategy

To solve the model in detrended recursive form, I perform value function iteration using a

discrete state-space for b and s = z, κ. The values for the policy functions, such as b′(b, s),

can lie outside the discrete grids. For this purpose, I use piecewise cubic-Hermite interpolation

polynomials (PCHIP), which solve the problems of using splines on monotonic functions, such

as the bond-price function, by incorporating its first derivatives.15

The discrete grid for b is constructed by setting 41 equally spaced points between the hy-

perbolic tangent of bmin and bmax, and then transforming back using the inverse-hyperbolic

tangent. This allows a greater number of points to be close to b = 0. The limits are set to

bmin = −0.75 and bmax = 1.5.

The discrete grid for z is constructed using the Tauchen (1986) method using 15 points and

an amplitude parameter equal to 3. The discrete grid of κ consists of only 2 points, as mentioned

before.

For ease of computation, I collapse the points z ∈ Z and κ ∈ κ into s ∈ S , where S = Z × κ

and s = (z, κ). The number of points will be 30. The transition probability of s, Pr(s′, s),

is a Kronecker product of Pr(κ′, κ) and Pr(z′, z), and allows calculations of expectations and

probabilities.

The stop rule for the value function iteration follows Chatterjee and Eyigüngör (2014), which

consists of iterating on Vo(b, s) and qB(b′, s) at the same time, until the criterion of convergence

falls below a tolerance of 10−6. The criterion of convergence is the maximum between that of

the value function and that of the sovereign portfolio price.

Once the terminal recursive formulation results are obtained, then I iterate backwards to

obtain the results with political-economy frictions, following the steps described in previous

sections. Using the initial recursive formulation policies, I proceed to simulate the values for

the variables of the system for 11,000 periods, using the last 10,000 to calculate the moments of

the model.

15 For a reference, see Fritsch and Carlson (1980), which provides the algorithm used in this chapter.

26

8 Results

In this section I show the results for several characterisations of the economy under study.

First, I review the mercantilist economy without political-economy frictions, and then I ad-

dress the effects of political-economy frictions. Next, I review the results for economies without

mercantilism, and the effect of political economy frictions for this group of economies. For

comparisons with the data, we refer to National Accounts data in Penn World Tables 8.1 for

emerging market and developing economies, which are the upper-middle, middle, and low in-

come ones as classified by the World Bank in the year 2000, to incorporate economies that now

have moved to the upper bracket of national income.

8.1 The mercantilist economy without political-economy frictions

The Table 3 show the moments of the model of the economy with mercantilism (ξ > 0) and

without political economy frictions (θ = 1), which will be our benchmark. This model yields

a volatility of consumption which is greater than that shown in the data, while the volatility

of gross domestic product is even greater, which is also the case for the trade balance. On the

other hand, the volatility of net foreign assets is lower than that of the data. In terms of cyclicity,

the trade balance is highly countercyclical compared to the data where it is mildly pro-cyclical,

while consumption and net foreign assets are less pro-cyclical than in the data.

Regarding average moments, this economy features a net public savings to GDP ratio of

30.1 percent, which would be in the 95-99 percentiles in the data, comparable to China during

the 2000s. The economy of this model never defaults, and suffers sudden stops in which its

tradable firms access public credit 0.8 percent of the time. This economy grows on average

3.6 percent a year, which is slightly higher than the average for an emerging market economy

according to data. When the government intervenes in the tradable sector it provides credit

equivalent to 9.9 percent of tradable output.

27

8.2 The political-economy frictions

The effect of the political-economy frictions is clearly depicted in Figure 4. These frictions

can explain differences in reserve accumulation of up to 20 percentage points of GDP when

the political-economy friction parameter θ is 40 percent or more. This is a vast difference in

reserve accumulation, comparable to the impact of an increase of one standard deviation in the

expropriation risk measure in the data. The effect on real GDP growth is milder, as an increase

of 40 percent in the political economy friction leads to a decrease of around 0.35 percentage

points in annual average growth, which is lower than the distribution shown in data. This

result is mostly due to the growth trend rate of knowledge coupled with a low elasticity of

imported inputs to growth, and persistent effects of lack of access to international credit.16

The average real exchange rate is mildly appreciated in the benchmark economy compared

to economies with political-economy frictions. For values of θ of 0.4 or above, the real exchange

rate is depreciated by almost 1.5 percent. Public credit to tradable firms in times of stress,

which is almost 10 percent of tradable output in the benchmark economy, falls to 7 percent at

the higher levels of the political-economy friction, which makes growth lower during periods

of sudden stops. This is due to the fact that the economy enters these episodes with lower

amounts of public savings to backstop the tradable firms.

It is important to highlight that within the mercantilist economy, further increases in θ bey-

ond 1.5 do not increase the differences from the benchmark economy, as the cost of not saving

becomes prohibitively high for a given level of the mercantilist parameters υ and ξ. As this

model does not engineer endogenous growth in the traditional sense, as there is no capital ac-

cumulation, it is very likely that the degree of mercantilism can be also affected by the degree

of political-economy frictions, an issue that remains unaddressed in this paper.

More information on the moments of economies with political-economy friction is shown

in Table 4. Two important features are that the volatility of macroeconomic aggregates and

16 For a review on these effects, refer to Gornemann (2014), which presents an endogenous growth model with agovernment that is more impatient than their households and can default on its debt. In theory, an economywith political-economy frictions is expected to show more sovereign defaults, and hence, much lower averagegrowth if the type of time-inconsistent preferences shown in this paper are present.

28

the number of periods under sudden stops decrease with the increase in political-economy

frictions. The former can be explained by the fact that during normal periods, the benchmark

economy achieves very high growth rates, increasing the volatility compared to the political-

economy friction economies; the latter is explained by the fact that demand for imported inputs

increases with the savings the government enters the period with, so the chances of having a

binding borrowing limit are higher for the benchmark economy than for the economies with

political economy frictions.

The next section will show the models with trend growth but no mercantilism, i.e., the

learning-by-trading parameter ξ = 0.

8.3 “Non-mercantilist” economies

It is important to compare the behaviour of an economy where there is no mercantilism

present, i.e. ξ = 0, and the trend growth rate of the economy is constant. As can be noticed

in Figure 5, the government now has net foreign liabilities. This is due to the fact that the eco-

nomy does not perceive important costs of not having enough savings to withstand sudden

stops. For an economy without political economy frictions, the difference in reserve accumu-

lation between mercantilist and non-mercantilist economies amounts to 46 percentage points

of GDP.17 Additionally, liabilities do not increase monotonically with the degree of political-

economy frictions. This stems from the fact that access to international capital markets is cur-

tailed more often than in the mercantilist economies, as shown in panel (c) of Figure 5. This

does not correspond to the relation in the data because PPG debt contains also aid flows and

official loans. In Alfaro et al. (2014), PPG debt with private lenders is an increasing function of

growth, which points out to the effect of access to capital markets in setting a limit to the public

borrowing abroad of the government.

The real exchange rate, shown in panel (b), also shows important differences between mer-

cantilist and non-mercantilist economies. The latter shows an appreciated level compared to

17 A pending exercise is to vary the degree of mercantilism, by using values of ξ ∈ (0, 0.1). This is expectedto change the amount of savings the public sector amasses. The range of public sector savings would fall inbetween the values shown in this paper for the mercantilist and non-mercantilist economies.

29

the former, which is consistent with the view that mercantilist economies should show more

depreciated exchange rates; and moreover, the relative appreciation is higher the higher the

political-economy frictions.

It is important to notices that political-economy frictions, as shown in panel (d), can explain

large differences in sovereign risk: a reduction of 70 percent in political-economy frictions can

reduce the sovereign spread by 800 basis points.

In Table 5 shows more detail regarding the moments of the model for different degrees of

political-economy frictions. In contrast with the mercantilist economy, the volatility of macroe-

conomic aggregates increases with higher political-economy frictions for low levels of θ, then

increases for higher levels of θ. As shown before, there is an increase in the number of periods

without access to international capital markets, due either to sudden stops or sovereign default.

9 Conclusions

This paper shows the role of mercantilism and sudden stops in emerging market economies

in explaining the existence of net positive international public assets, while political economy

frictions can account for the varying degrees of asset accumulation across economies with oth-

erwise similar characteristics. An increase of 50 percent in these frictions implies a reduction

of almost 20 percentage points of GDP in net foreign public assets, in the context of economies

with growth externalities, i.e. mercantilism. On the other hand, if mercantilism is not taken

into account, sovereigns accumulate net foreign liabilities, which dampens their access to inter-

national markets. For this reason, political-economy frictions explain differences in net foreign

positions to a lesser extent (5 percentage points of GDP), but can explain differences in sov-

ereign risk: a reduction of 70 percent in political-economy frictions can reduce the sovereign

spread by 800 basis points.

As far as I know, this is the first paper providing a plausible framework for a small open

economy where the government builds a positive net foreign asset position, while stressing the

important differences that political-economy frictions and mercantilist strategies make between

the net foreign-asset positions of the governments in emerging-market and developing econom-

30

ies. This is important to consider in the perspective of global imbalances: the pursuit of growth

externalities in tradable sectors and the improvement in institutional quality in emerging-market

and developing markets may exacerbate the savings-glut problem in the world economy.

Further extensions may need to be considered to improve on the explanation of GDP growth

outcomes. As mentioned before, it may be necessary to embed models of endogenous growth,

such as growth due to increased varieties of tradable goods, and add capital accumulation. This

way, growth trends will depend to a greater extent on institutional quality, as a vast literature

has shown.

On the other hand, it may be important to study the accumulation of public net foreign asset

positions under the lens of monetary and exchange rate arrangements. Recently, small open

economies such as Denmark and Switzerland have amassed vast amounts of foreign reserves

under pressure during the European Debt Crisis, and many economists have pointed out the

use of foreign reserves in order to control exchange rate movements finally, one could explore

the use of foreign reserves in mitigating possible banking and currency crises in emerging-

market and developing economies.

References

Acemoglu, Daron, James A. Robinson, and Simon Johnson, “The Colonial Origins of Compar-ative Development: An Empirical Investigation,” The American Economic Review, December2001, 91 (5), 1369–1401.

, Suresh Naidu, Pascual Restrepo, and James A. Robinson, “Democracy Does CauseGrowth,” Working Paper 20004, National Bureau of Economic Research March 2014.

Aguiar, Mark and Manuel Amador, “Growth in the Shadow of Expropriation,” The QuarterlyJournal of Economics, 2011, 126 (2), 651–697.

and , “Sovereign Debt,” in Elhanan Helpman, Kenneth Rogoff, and Gita Gopinath, eds.,Handbook of International Economics, Vol. 4 of Handbooks in Economics, Oxford, United King-dom: Elsevier, 2014, chapter 11, pp. 647–688.

, , and Gita Gopinath, “Investment Cycles and Sovereign Debt Overhang,” Review of Eco-nomic Studies, 01 2009, 76 (1), 1–31.

Aizenman, Joshua and Jaewoo Lee, “International Reserves: Precautionary Versus MercantilistViews, Theory and Evidence,” Open Economies Review, 2007, 18 (2), 191–214.

31

and Nancy Marion, “International Reserve Holdings With Sovereign Risk and Costly TaxCollection,” Economic Journal, 07 2004, 114 (497), 569–591.

Alesina, Alberto and David Dollar, “Who Gives Foreign Aid to Whom and Why?,” Journal ofEconomic Growth, 2000, 5 (1), 33–63.

Alfaro, Laura and Fabio Kanczuk, “Optimal Reserve Management and Sovereign Debt,”Journal of International Economics, 2009, 77 (1), 23–36.

and , “Carry Trade, Reserve Accumulation, and Exchange-Rate Regimes,” Working Paper19089, National Bureau of Economic Research June 2013.

, Sebnem Kalemli-Özcan, and Vadym Volosovych, “Why Doesn’t Capital Flow From Richto Poor Countries? an Empirical Investigation,” The Review of Economics and Statistics, 2008,90 (2), 347–368.

, , and , “Sovereigns, Upstream Capital Flows and Global Imbalances,” Journal of theEuropean Economic Association, October 2014, 12 (5), 1240–1248.

Amador, Manuel, “A Political Economy Model of Sovereign Debt Repayment.” PhD disserta-tion, Stanford Graduate School of Business September 2003.

, “Sovereign Debt and the Tragedy of Commons,” Mimeo 2012.

Angeletos, George-Marios and Vasia Panousi, “Financial Integration, Entrepreneurial Riskand Global Dynamics,” Journal of Economic Theory, 2011, 146 (3), 863–896.

Arellano, Cristina, “Default Risk and Income Fluctuations in Emerging Economies,” The Amer-ican Economic Review, 2008, 98 (3), 690–712.

Azzimonti, Marina and Matthew Talbert, “Polarized Business Cycles,” Journal of MonetaryEconomics, 2014, 67, 47 – 61.

Baer, Werner, “Industrialization in Latin America: Successes and Failures,” The Journal of Eco-nomic Education, 1984, 15 (2), 124–135.

Beck, Thorsten, George Clarke, Alberto Groff, Philip Keefer, and Patrick Walsh, “New Toolsin Comparative Political Economy: The Database of Political Institutions,” The World BankEconomic Review, 2001, 15 (1), 165–176.

Benigno, Gianluca and Luca Fornaro, “Reserve Accumulation, Growth and Financial Crisis,”Working Paper, London School of Economics July 2012.

Bernanke, Ben S, “The Global Saving Glut and the U.S. Current Account Deficit,” Remarksby Chairman Ben S. Bernanke at the Sandridge Lecture, Virginia Association of Economists,Richmond, Virginia, March 10th. 2005.

Bianchi, Javier, Juan Carlos Hatchondo, and Leonardo Martínez, “International Reserves andRollover Risk,” Working Paper 18628, National Bureau of Economic Research December 2012.

Buera, Francisco J. and Yongseok Shin, “Productivity Growth and Capital Flows: The Dy-namics of Reforms,” Working Paper 15268, National Bureau of Economic Research August2009.

32

Caballero, Ricardo J and Stavros Panageas, “Hedging Sudden Stops and Precautionary Con-tractions,” Journal of Development Economics, 2008, 85 (1), 28–57.

, Emmanuel Farhi, and Pierre-Olivier Gourinchas, “An Equilibrium Model of ‘Global Im-balances’ and Low Interest Rates,” American Economic Review, 2008, 98 (1), 358–393.

Catão, Luis, “The Failure of Export-Led Growth in Brazil and México, C. 1870-1930,” ResearchPapers 31, University of London Institute for the Study of the Americas 1992.

Chatterjee, Satyajit and Burcu Eyigüngör, “Continuous Markov Equilibria With Quasi-Geometric Discounting,” Working Paper 14–6, Research Department, Federal Reserve Bankof Philadelphia 2014.

Chinn, Menzie and Hiro Ito, “A New Measure of Financial Openness,” Journal of ComparativePolicy Analysis, September 2008, 10 (3), 309–322.

Chinn, Menzie D, Barry Eichengreen, and Hiro Ito, “A Forensic Analysis of Global Imbal-ances,” Oxford Economic Papers, 2014, 66 (2), 465–490.

Choi, Woon Gyu, Sunil Sharma, and Maria Strömqvist, “Capital Flows, Financial Integra-tion, and International Reserve Holdings: The Recent Experience of Emerging Markets andAdvanced Economies,” Working Paper 151, International Monetary Fund. July 2007.

Cortés-Conde, Roberto, “Export-Led Growth in Latin America: 1870-1930,” Journal of LatinAmerican Studies, 1992, 24, 163–179.

D’Erasmo, Pablo N., “Essays in Dynamic Macroeconomics.” PhD dissertation, University ofTexas at Austin 2008.

Dooley, Michael, David Folkerts-Landau, and Peter Garber, “The Revived Bretton WoodsSystem,” International Journal of Finance and Economics, 2004, 9 (4), 307–313.

Durdu, Ceyhun Bora, Enrique G. Mendoza, and Marco E. Terrones, “Precautionary Demandfor Foreign Assets in Sudden Stop Economies: An Assessment of the New Mercantilism,”Journal of Development Economics, 2009, 89 (2), 194–209.

Eaton, Jonathan and Mark Gersovitz, “Debt With Potential Repudiation: Theoretical and Em-pirical Analysis,” The Review of Economic Studies, 1981, 48 (2), 289–309.

Eichengreen, Barry, Golden Fetters: The Gold Standard and the Great Depression, 1919-1939, OxfordUniversity Press, 1992.

, “Global Imbalances and the Lessons of Bretton Woods,” Working Paper 10497, NationalBureau of Economic Research May 2004.

, “Global Imbalances: The New Economy, the Dark Matter, the Savvy Investor, and the Stand-ard Analysis,” Journal of Policy Modeling, September 2006, 28 (6), 645–652.

and Yung Chul Park, “Global Imbalances and Emerging Markets,” Paper Prepared for theConference on European and Asian Perspectives on Global Imbalances in Beijing, July 13th-14th July 2006.

33

Ekelund, Robert B. Jr. and Robert F. Hébert, A History of Economic Theory and Method, WavelandPress, 2013.

Feenstra, Robert C., Robert Inklaar, and Marcel P. Timmer, “The Next Generation of the PennWorld Table,” American Economic Review, October 2015, 105 (10), 3150–3182.

Frankel, Jeffrey, Carlos Végh, and Guillermo Vuletin, “On Graduation From Fiscal Procyclic-ality,” Journal of Development Economics, 2013, 100 (1), 32–47.

Fritsch, Frederick N and Ralph E Carlson, “Monotone Piecewise Cubic Interpolation,” SIAMJournal on Numerical Analysis, 1980, 17 (2), 238–246.

Gertler, Mark and Peter Karadi, “A Model of Unconventional Monetary Policy,” Journal ofMonetary Economics, 2011, 58 (1), 17–34.

Gómez-Galvarriato, Aurora and Jeffrey G. Williamson, “Was It Prices, Productivity or Policy?Latin American Industrialisation After 1870,” Journal of Latin American Studies, 2009, 41 (4),663–694.

Gornemann, Nils, “Sovereign Default, Private Investment, and Economic Growth,” Mimeo2014.

Gourinchas, Pierre-Olivier and Hélène Rey, “External Adjustment, Global Imbalances andValuation Effects,” in Elhanan Helpman, Kenneth Rogoff, and Gita Gopinath, eds., Handbookof International Economics, Vol. 4 of Handbooks in Economics, Oxford, United Kingdom: Elsevier,2014, chapter 10, pp. 585–645.

and Olivier Jeanne, “Capital Flows to Developing Countries: The Allocation Puzzle,” TheReview of Economic Studies, 2013, 80 (4), 1484–1515.

Gourio, François, Michael Simer, and Adrien Verdelhan, “Uncertainty and International Cap-ital Flows,” Mimeo, Federal Reserve Bank of Chicago 2015.

Greenspan, Alan, “Testimony of Chairman Alan Greenspan,” Federal Reserve Board’s SemiannualMonetary Policy Report to the Congress Before the Committee on Banking, Housing, and UrbanAffairs, U.S. Senate., Federal Reserve Board February 16th, 2005.

Jeanne, Olivier and Romain Rancière, “The Optimal Level of International Reserves for Emer-ging Market Countries: A New Formula and Some Applications*,” The Economic Journal,2011, 121 (555), 905–930.

Jones, Benjamin F. and Benjamin A. Olken, “Do Leaders Matter? National Leadership andGrowth Since World War II,” The Quarterly Journal of Economics, 2005, 120 (3), 835–864.

and , “Hit or Miss? the Effect of Assassinations on Institutions and War,” American Eco-nomic Journal: Macroeconomics, July 2009, 1 (2), 55–78.

Keefer, Philip and David Stasavage, “The Limits of Delegation: Veto Players, Central Bank In-dependence, and the Credibility of Monetary Policy,” American Political Science Review, 2003,97 (3), 407–423.

34

Korinek, Anton and Luis Servén, “Undervaluation Through Foreign Reserve Accumulation:Static Losses, Dynamic Gains,” Policy Research Working Paper 5250, The World Bank March2010.

Krishnamurthy, Arvind and Annette Vissing-Jørgensen, “The Demand for Treasury Debt,”Working Paper 12881, National Bureau of Economic Research January 2007.

Laibson, David, “Golden Eggs and Hyperbolic Discounting,” The Quarterly Journal of Econom-ics, 1997, 112 (2), 443–477.

Lindqvist, Erik and Robert Östling, “Political Polarization and the Size of Government,” Amer-ican Political Science Review, 2010, 104 (3), 543–565.

Mendoza, Enrique G. and Vivian Z. Yue, “A General Equilibrium Model of Sovereign Defaultand Business Cycles,” The Quarterly Journal of Economics, 2012, 27 (2), 889–946.

, Vincenzo Quadrini, and Jose-Víctor Ríos-Rull, “Financial Integration, Financial Develop-ment, and Global Imbalances,” Journal of Political Economy, 2009, 117 (3), 371–416.

Obstfeld, Maurice and Alan M. Taylor, “Sovereign Risk, Credibility and the Gold Standard:1870-1913 Versus 1925-31,” The Economic Journal, 2003, 113 (487), 241–275.

Papaioannou, Elías, “What Drives International Financial Flows? Politics, Institutions andOther Determinants,” Journal of Development Economics, 2009, 88 (2), 269–281.

Prebisch, Raúl and Gustavo Martínez-Cabañas, “El Desarrollo Económico De América LatinaY Sus Principales Problemas,” El Trimestre Económico, 1949, 16 (63), 347–431.

Reinhart, Carmen M, Vincent Reinhart, and Takeshi Tashiro, “Does Reserve AccumulationCrowd Out Investment?,” Journal of International Money and Finance, 2016, 63, 89–111.

Rodrik, Dani, “The Real Exchange Rate and Economic Growth,” Brookings Papers on EconomicActivity, 2008, 2008 (2), 365–412.

Schmitt-Grohé, Stephanie and Martín Uribe, “Is Optimal Capital-Control Policy Countercyc-lical in Open-Economy Models With Collateral Constraints?,” Working Paper 22481, NationalBureau of Economic Research August 2016.

Tauchen, George, “Finite State Markov-Chain Approximations to Univariate and VectorAutoregressions,” Economics Letters, 1986, 20 (2), 177–181.

Tomz, Michael and Mark L. J. Wright, “Sovereign Theft: Theory and Evidence About Sover-eign Default and Expropriation,” in William Hogan and Federico Sturzenegger, eds., The Nat-ural Resources Trap: Private Investment Without Public Commitment, MIT Press, 2010, chapter 3,pp. 69–118.

Woodford, Michael, “Is an Undervalued Currency the Key to Economic Growth?,” Departmentof Economics Discussion Papers 0809-13, Columbia University January 2009.

Wright, Mark L.J., “The Theory of Sovereign Debt and Default,” in Gerard Caprio, ed., Encyc-lopedia of Financial Globalization, Oxford, United Kingdom: Elsevier, forthcoming.

35

Appendix A

In this appendix, I provide further details in the solution of the benchmark model, with

mercantilism, sovereign default and sudden stops.

A.1 Competitive equilibrium

To solve the competitive equilibrium for the private sector, I use the detrended first order

conditions of the households and firms to obtain an equilibrium equation as a function of labour

in the tradable sector, taking as giving the net foreign asset decisions of the government.

A.1.1 Unconstrained solution

Starting from the first order conditions of the tradable sector firms’, we can find an expres-

sion of imported inputs as a function of parameters and tradable labour:

mt = χ1α LT

t (60)

where χt =[(1− α) zt/PM

t]. Using this expression, we can find an expression for the equilib-

rium wage of the economy:

w∗t = αztχ1α−1t (61)

Using the first order condition of nontradable sector firms’ and that of households, plus the

resource constraint and the nontradable market clearing condition, I arrive to an expression to

calculate the equilibrium labour in the tradable sector:

F(

LTt

)=

[aw∗t

(1− a)γ

] 11+ν (

1− LTt

) 1+γν1+ν − w∗t LT

t − bt + qt

[υ + χ

ξαt

(LT

t

)ξ]

bt+1 (62)

This equation has the possibility of having more than one root in the interval [0, 1]. For this

matter, the algorithm to solve the equation uses 41 Chebyshev collocation points, and brackets

the function for the highest root, as to keep the monotonicity of the relation between tradable

36

labour and next-period sovereign net foreign assets.

Once a solution for labour in tradable sector is calculated, using the resource constraint

and the nontradable market clearing conditions, is straightforward to calculate the rest of the

variables.

A.1.2 Unconstrained solution under sovereign default

This case is directly solvable by using bt = bt+1 = 0 allows for a straightforward solution:

F(

LTt

)=

[aw∗t

(1− a)γ

] 11+ν (

1− LTt

) 1+γν1+ν − w∗t LT

t (63)

It is important to highlight that for a sovereign default, the firms are subject to sudden stops

with the same parameter κ under no sovereign default, i.e. κ = 0.12, and there is an symmetrical

loss of 10% in tradable and nontradable consumption units.

A.1.3 Constrained solution

When the realisation of borrowing limit κ is a low level, we need to check whether φPMmut >

κt, where mut is the unconstrained level of imported inputs. If the sovereign has enough re-

sources, it could finance the gap and restore the first-best equilibrium, or get closer to it:

mt =κ + max [min (φPMmu

t − κt, (1− ψ) bt) , 0]φPM

(64)

Even if the government can help achieve mut , there is a social cost for such credits which im-

pacts transfers to households. Hence, we need to recalculate the level of labour in the tradable

sector under this situation, by solving the following:

F(

LTt

)= aαzt

(LT

t

)α−1m1−α

t − (1− a) γ[αzt

(LT

t

)αm1−α

t

+bt −ψ

1− ψdt

]1+ν (1− LT

t

)−γν−1 (65)

In the case of sovereign default, the firms are always subject to a borrowing limit of κ = 0.12.

37

Since they default on debt, cannot access debt and cannot provide credit to tradable firms, the

calculation is much simpler by substituting bt = dt = 0 into the previous equation.

38

AGO

ALB

ARG

ARM AZEBFABGD BGRBLR

BOL

BRA

BWA

CHLCHN

CIV

CMR

COD

COG

COLCRI

CZE

DOMECU EGY

EST

ETHGAB

GHA

GINGMB

GNB

GTM

HND

HRVHUNIDN

INDIRN

JAMJOR

KAZ

KEN

KORLBN

LBR

LKALTULVA

MAR

MDA

MDG

MEX

MLI

MLT

MNG

MOZ

MWI

MYS

NER

NGA

OMNPAK

PAN

PERPHL POLPRYROURUS

SDNSEN

SGP

SLE

SLV

SRBSVK SVN

SYR

TGO

THATTO

TUNTUR

TWN

TZA

UGA

UKR

URYVEN VNM

YEM

ZAF

ZMB

ZWE

-1.5

-1

-.5

0

.5

1

Net F

orei

gn P

ublic

Ass

ets

(% o

f GDP

)

3 5 7 9 11Expropriation Risk [0-12]

Figure 1: Correlation between Net Public Foreign Assets and Expropriation Risk.Plot shows average real reserves stock as share of real GDP against the averagelevel of the investment profile subcomponent of the ICRG political risk index, whichmeasures Expropriation Risk. In the figure, a high value in the Expropriation Riskmeasure implies a low expropriation risk. Averages are taken over available databetween 1980 and 2011. Source: Own elaboration on WDI Database Archive andICRG.

39

AGOALBARG ARM

AUSAUTAZE BELBFA

BGD

BGR

BHR

BHS

BLR

BOLBRA BRN

BWA

CAN

CHE

CHLCHN

CIV CMRCOD

COGCOL CRI

CYPCZE

DEU

DNK

DOMECUEGY ESP

EST

ETH

FINFRAGAB GBRGHA

GIN

GMBGNB GRCGTM

HKG

HND

HRVHUN

IDNIND

IRLIRN

IRQ

ISLISR

ITA

JAM

JOR

JPNKAZ

KEN

KOR

KWT

LBN

LBR LKA

LTU

LUX

LVAMARMDA

MDG MEXMLI

MLT

MNGMOZ

MWI

MYS

NAMNER

NGA

NLD

NOR

NZL

OMN

PAKPAN

PERPHL POLPRT

PRYQATROU

RUS

SAU

SDNSEN

SGP

SLE

SLV

SRB

SURSVK SVN

SWESYR

TGO THA

TTO

TUNTUR

TWN

TZA UGAUKRURY

USA

VEN

VNM

YEM

ZAFZMBZWE0

.2

.4

.6

.8

Inte

rnat

iona

l Res

erve

s (%

of G

DP)

3 5 7 9 11Expropriation Risk [0-12]

(a) International Reserves

AGO

ALB

ARG

ARMAZE

BFABGD BGR

BLR

BOL

BRA BWACHLCHN

CIV

CMR

COD

COG

COL

CRI

CZE

DOM

ECU EGY

EST

ETH

GAB

GHA

GIN GMB

GNB

GTM

HND

HRVHUN

IDNINDIRN

JAMJOR

KAZ

KENKOR

LBN

LBR

LKALTULVA

MAR

MDA

MDG

MEX

MLI

MLTMNG

MOZ

MWI

MYS

NER

NGA

OMNPAK

PAN

PERPHL POLPRY

ROURUS

SDNSEN

SGP

SLE

SLV

SRB

SVKSVN

SYR

TGO

THA

TTOTUN

TUR

TWN

TZA

UGA

UKR

URYVENVNM

YEM

ZAF

ZMB

ZWE

-.5

0

.5

1

1.5

PPG

Deb

t (%

of G

DP)

3 5 7 9 11Expropriation Risk [0-12]

(b) Public and Publicly Guaranteed Debt

Figure 2: Correlation between Foreign Reserves and PPG Debt, and ExpropriationRisk. In the top panel, the plot shows average real reserves stock as share of realGDP against the average level of expropriation risk, as measured by the investmentprofile variable in ICRG. In the bottom panel, the plot shows average real PPG debtas share of real GDP against the average level of expropriation risk. Averages aretaken over available data between 1980 and 2011. Source: Own elaboration on WDIDatabase Archive and ICRG.

40

0 1 2 3 4 5 6 7 8 9 100.60

0.65

0.70

0.75

0.80

0.85

0.90

0.95

Periods

Inte

rtem

pora

ldis

coun

trat

e

δ=0δ=0.25δ=0.50δ=0.70

Figure 3: Quasi-hyperbolic discounting. The figure shows the one-period discountfunction for different horizons, given β = 0.9, and θ = 1.5. Traditional models fea-ture θ = 1 and δ = 0, so there are no political-economy frictions. If θ increases to 1.5as in this figure, the discount rate between periods 0 and 1 is lower than the standardcase, but in the following period is the same as in a benevolent government. Thisalso the case under standard quasi-hyperbolic discounting à la Laibson (1997), dueto the fact that the distortions are not expected to last due to lack of incumbencyadvantage, i.e. δ = 0. If incumbency advantage is higher (recall δ ∈ [0, 1]), thedifference in discounting is less abrupt than in the case of δ = 0, but persists longer.

41

1 1.1 1.2 1.3 1.4 1.50.100

0.150

0.200

0.250

0.300

0.350

θ

1 1.1 1.2 1.3 1.4 1.50.032

0.033

0.034

0.035

0.036

θ

(a) Net foreign assets / GDP (b) Real GDP growth

1 1.1 1.2 1.3 1.4 1.50.958

0.960

0.962

0.964

0.966

0.968

0.970

0.972

0.974

θ

1 1.1 1.2 1.3 1.4 1.50.065

0.070

0.075

0.080

0.085

0.090

0.095

0.100

θ

(c) Real exchange rate (d) Public credit / Tradable Output

Figure 4: The effects of political-economy frictions θ on key variables of the mercant-ilist economy. Panel (a) shows the average net foreign assets over GDP accumulatedby the government during normal times. Panel (b) shows the average real growthof GDP during all periods. Panel (c) shows the average real exchange rate during allperiods. Panel (d) shows the public credit given to the firms in the tradable sector bythe government, scaled by the tradable output, during low borrowing limit shocks.

42

1 1.5 2 2.5 3 3.5−0.220

−0.210

−0.200

−0.190

−0.180

−0.170

−0.160

θ

1 1.5 2 2.5 3 3.5

1.020

1.040

1.060

1.080

1.100

1.120

θ

(a) Net foreign assets / GDP (b) Real exchange rate

1 1.5 2 2.5 3 3.50.000

0.020

0.040

0.060

0.080

0.100

0.120

0.140

θ

1 1.5 2 2.5 3 3.50.000

0.020

0.040

0.060

0.080

θ

(c) Lack of access to capital markets (d) Sovereign spread

Figure 5: The effects of political-economy frictions θ on key variables of the non-mercantilist economy. Panel (a) shows the average net foreign assets over GDP ac-cumulated by the government during normal times. Panel (b) shows the averagereal exchange rate during all periods. Panel (c) shows the share of periods that theeconomy spends under sudden stops and sovereign default. Panel (d) shows theaverage sovereign spread over the foreign sovereign bonds.

43

Tabl

e1:

Res

ults

ofG

ouri

ncha

san

dJe

anne

(201

3)co

ntro

lling

for

expr

opri

atio

nri

sk(1

)(2

)(3

)(4

)(5

)(6

)(7

)(8

)(9

)(1

0)(1

1)(1

2)Pu

blic

flow

sPP

Gde

btR

eser

ves

Publ

icflo

ws

PPG

debt

Res

erve

sPu

blic

flow

sPP

Gde

btR

eser

ves

Prod

ucti

vity

catc

h-up

(π)

-0.8

4***

-1.1

8***

-0.0

4-1

.15*

**-0

.78*

**-1

.12*

**-0

.04

-1.0

8***

1.08

0.70

0.16

0.54

(0.1

9)(0

.19)

(0.1

0)(0

.15)

(0.2

1)(0

.21)

(0.1

0)(0

.17)

(0.8

3)(0

.74)

(0.3

8)(0

.59)

Init

ialc

apit

alab

unda

nce

-0.1

8+-0

.11

-0.0

9+-0

.03

-0.1

3-0

.07

-0.0

8+0.

02-0

.06

-0.0

1-0

.08

0.07

(0.1

0)(0

.09)

(0.0

5)(0

.08)

(0.1

1)(0

.10)

(0.0

5)(0

.08)

(0.1

1)(0

.10)

(0.0

5)(0

.08)

Init

iald

ebt

0.00

-0.0

0-0

.00

-0.0

0-0

.00

-0.0

0-0

.00

-0.0

0-0

.00

-0.0

1-0

.00

-0.0

0+(0

.00)

(0.0

0)(0

.00)

(0.0

0)(0

.00)

(0.0

0)(0

.00)

(0.0

0)(0

.00)

(0.0

0)(0

.00)

(0.0

0)Po

pula

tion

grow

th-0

.21*

-0.1

5+0.

07+

-0.2

2**

-0.2

0*-0

.14+

0.07

+-0

.21*

*-0

.22*

-0.1

6*0.

07-0

.23*

**(0

.09)

(0.0

8)(0

.04)

(0.0

6)(0

.09)

(0.0

8)(0

.04)

(0.0

7)(0

.09)

(0.0

8)(0

.04)

(0.0

6)O

penn

ess

(Chi

nn-I

to)

-0.1

6*-0

.13*

0.03

-0.1

6***

-0.1

5*-0

.12*

0.03

-0.1

6***

-0.1

4*-0

.11*

0.03

-0.1

5***

(0.0

6)(0

.05)

(0.0

3)(0

.04)

(0.0

6)(0

.06)

(0.0

3)(0

.04)

(0.0

6)(0

.05)

(0.0

3)(0

.04)

Ope

nnes

s×

π-0

.69*

**-0

.00

-0.6

9***

-0.6

6***

0.00

-0.6

6***

-0.6

5***

0.01

-0.6

5***

(0.1

7)(0

.09)

(0.1

4)(0

.18)

(0.0

9)(0

.14)

(0.1

7)(0

.09)

(0.1

4)In

vest

men

tpro

file

-0.1

0-0

.08

0.00

-0.0

8+-0

.10+

-0.0

8-0

.00

-0.0

8+(0

.06)

(0.0

6)(0

.03)

(0.0

4)(0

.06)

(0.0

5)(0

.03)

(0.0

4)in

vest

men

tpro

file×

π-0

.30*

-0.3

0*-0

.03

-0.2

6**

(0.1

3)(0

.12)

(0.0

6)(0

.09)

Inte

rcep

t0.

67*

0.50

*0.

210.

301.

18**

0.89

*0.

200.

69*

1.25

**0.

96*

0.21

0.76

*(0

.27)

(0.2

4)(0

.13)

(0.2

0)(0

.41)

(0.3

8)(0

.18)

(0.3

0)(0

.40)

(0.3

6)(0

.18)

-0.2

8

Num

ber

ofob

serv

atio

ns62

6262

6257

5757

5757

5757

57A

djus

ted-

R2

0.37

0.50

0.11

0.59

0.40

0.52

0.10

0.62

0.45

0.57

0.09

0.67

Stan

dard

erro

rsin

pare

nthe

ses.∗ p

<0.

05,∗∗

p<

0.01

,∗∗∗

p<

0.00

1.N

otes

:All

data

ista

ken

from

Gou

rinc

has

and

Jean

ne(2

013)

,exc

eptf

orEx

prop

riat

ion

Ris

k,w

hich

isth

ein

vest

men

tpro

file

subc

ompo

nent

ofIC

RG

inde

x,av

erag

edov

er19

84-2

000.

The

depe

nden

tvar

iabl

esar

eth

eflo

ws

ofth

em

enti

oned

cate

gory

scal

edby

GD

P.π

ispr

oduc

tivi

tyca

tch-

upw

ith

Uni

ted

Stat

es,(

k/y)

0is

the

init

ialc

apit

alra

tio,

(d/

y)0

isth

ein

itia

lsto

ckof

asse

ts/d

ebt,

nis

the

aver

age

grow

thof

popu

lati

onbe

twee

n15

-64

year

sol

d.K

AO

PE

Nis

the

capi

talo

penn

ess

from

Chi

nnan

dIt

o(2

008)

.

44

Table 2: Main parameters

Parameter Value Description Source/Target

σ 2 CRRA coefficient Standard valuesa 0.31 Share of tradables in consumption Schmitt-Grohé and Uribe (2016)ζ 0.2 CES parameter Schmitt-Grohé and Uribe (2016)β 0.90 Intertemporal discount factor Standard valuesr∗ 0.04 International interest rate (%, annual) Standard valuesη 1/3 Probability of redemption Bianchi et al. (2012)α 0.6 Share of labour in tradable output Standard valuesγ 0.6 Share of labour in nontradable output Standard values

PM 1 Price of imported inputs Benigno and Fornaro (2012)ψ 0.5 Share of public FC loans lost Benigno and Fornaro (2012)φ 0.7 Share of foreign-financed imported inputs Mendoza and Yue (2012)ξ 0.1 Elasticity of knowledge to imports Benigno and Fornaro (2012)υ 0.17 Trend growth rate of knowledge Average annual growth of 3.4%.θ 1-1.5 Political-economy friction Aguiar and Amador (2011)δ 0 Incumbent advantage Aguiar and Amador (2011)

κL 0.12 Low value of κ Trend growth rate of -2.5%

45

Table 3: Results for mercantilist economy without political-economy frictions

Model1 Data1,2

Standard Correlation Standard Correlationdeviation with GDP deviation with GDP

Tradable output 0.10 -0.65 ... ...Tradable consumption 0.13 0.93 ... ...Nontradable consumption 0.05 0.83 ... ...Consumption 0.08 0.90 0.07 0.40GDP 0.07 1.00 0.04 1.00Imported inputs 0.07 0.43 ... ...Wages 0.07 0.99 ... ...Net foreign assets/GDP 0.18 0.15 0.40 0.09Trade Balance / GDP 0.07 -0.80 0.23 0.03Domestic interest rate 0.00 0.00 ... ...Real exchange rate 0.15 0.81 ... ...

Number of episodes

Sovereign Default 0Sudden stops3 823

Average statistics ( percent)

Public savings/GDP4 30.10GDP growth5 3.60Real exchange rate 97.30Public credit/tradable output6 9.9

1 Series are detrended using HP filter with parameter λ = 6.25, and statistics are calculatedusing the deviations from the trend.2 Uses national accounts data from Penn World Tables 8.1, for countries classified as upper-middle, middle, and low income, according to the World Bank Organisation for the year2000.3 Includes only those episodes where the borrowing limit is binding.4 Compared over periods when there is full access to international capital markets.5 Real GDP is computed deflating nominal GDP by the consumer price index of the eco-nomy.6 Compared over episodes where the borrowing limit has a low realisation.

46

Tabl

e4:

Res

ults

for

mer

cant

ilist

econ

omy

wit

hpo

litic

al-e

cono

my

fric

tion

s

θ=

1θ=

1.1

θ=

1.2

θ=

1.3

θ=

1.4

θ=

1.5

σx

σx,

GD

Pσ

xσ

x,G

DP

σx

σx,

GD

Pσ

xσ

x,G

DP

σx

σx,

GD

Pσ

xσ

x,G

DP

Trad

able

outp

ut0.

10-0

.65

0.08

-0.4

60.

07-0

.27

0.05

0.06

0.05

0.39

0.05

0.41

Trad

able

cons

umpt

ion

0.13

0.93

0.11

0.90

0.09

0.86

0.07

0.82

0.05

0.72

0.05

0.72

Non

trad

able

cons

umpt

ion

0.05

0.83

0.04

0.77

0.04

0.69

0.03

0.59

0.02

0.37

0.02

0.36

Con

sum

ptio

n0.

080.

900.

070.

850.

060.

790.

040.

720.

030.

570.

030.

56G

DP

0.07

1.00

0.06

1.00

0.05

1.00

0.04

1.00

0.03

1.00

0.03

1.00

Impo

rted

inpu

ts0.

070.

430.

070.

490.

070.

570.

070.

700.

070.

770.

070.

78W

ages

0.07

0.99

0.05

0.99

0.05

0.98

0.04

0.98

0.03

0.97

0.03

0.97

Net

fore

ign

asse

ts/G

DP

0.18

0.15

0.14

0.20

0.11

0.24

0.07

0.28

0.05

0.31

0.05

0.30

Trad

eBa

lanc

e/

GD

P0.

07-0

.80

0.06

-0.7

40.

05-0

.68

0.04

-0.5

60.

03-0

.37

0.03

-0.3

5D

omes

tic

inte

rest

rate

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

Rea

lexc

hang

era

te0.

150.

810.

120.

730.

110.

660.

090.

540.

080.

440.

080.

43

Num

ber

ofep

isod

es

Sove

reig

nde

faul

t0

00

00

0Su

dden

stop

s182

382

382

077

076

375

6

Ave

rage

stat

isti

cs(%

)

Publ

icsa

ving

s/G

DP2

30.3

24.7

20.5

15.0

11.0

10.7

Rea

lGD

Pgr

owth

33.

63.

53.

43.

33.

33.

2R

eale

xcha

nge

rate

97.3

96.8

96.6

96.2

96.0

96.0

Publ

iccr

edit

/tra

dabl

eou

tput

49.

98.

88.

17.

46.

86.

8

Seri

esar

ede

tren

ded

usin

gH

Pfil

ter

wit

hpa

ram

eter

λ=

6.25

,so

stat

isti

csar

eca

lcul

ated

usin

gth

ede

viat

ions

from

the

tren

d.1

Con

side

rsth

ose

epis

odes

whe

reth

ebo

rrow

ing

limit

isbi

ndin

g.2

Con

side

rsth

epe

riod

sw

hen

ther

eis

full

acce

ssto

inte

rnat

iona

lcap

ital

mar

kets

.3

For

this

stat

isti

c,th

eG

DP

isde

flate

dby

the

cons

umer

pric

ein

dex

ofth

eec

onom

y.4

Con

side

rsth

ose

epis

odes

whe

reth

ebo

rrow

ing

limit

has

alo

wre

alis

atio

n.

47

Tabl

e5:

Res

ults

for

non-

mer

cant

ilist

econ

omy

wit

hpo

litic

al-e

cono

my

fric

tion

s

θ=

1θ=

1.5

θ=

2θ=

2.5

θ=

3θ=

3.5

σx

σx,

GD

Pσ

xσ

x,G

DP

σx

σx,

GD

Pσ

xσ

x,G

DP

σx

σx,

GD

Pσ

xσ

x,G

DP

Trad

able

outp

ut0.

04-0

.48

0.05

-0.5

80.

07-0

.58

0.06

-0.1

40.

060.

180.

060.

29Tr

adab

leco

nsum

ptio

n0.

130.

960.

160.

960.

180.

960.

140.

900.

100.

860.

090.

85N

ontr

adab

leco

nsum

ptio

n0.

050.

890.

070.

900.

080.

820.

090.

110.

09-0

.31

0.10

-0.4

0C

onsu

mpt

ion

0.08

0.94

0.10

0.94

0.11

0.91

0.09

0.53

0.08

0.12

0.07

-0.0

0G

DP

0.06

1.00

0.07

1.00

0.07

1.00

0.07

1.00

0.06

1.00

0.06

1.00

Impo

rted

inpu

ts0.

060.

890.

060.

830.

060.

730.

060.

450.

070.

130.

070.

08W

ages

0.05

0.99

0.06

0.99

0.06

0.95

0.06

0.39

0.06

-0.0

80.

06-0

.19

Net

fore

ign

asse

ts/G

DP

0.07

0.32

0.08

0.34

0.09

0.33

0.09

0.05

0.09

-0.1

90.

09-0

.25

Trad

eBa

lanc

e/

GD

P0.

07-0

.89

0.08

-0.9

00.

10-0

.89

0.07

-0.6

80.

05-0

.43

0.05

-0.3

4D

omes

tic

inte

rest

rate

0.00

0.00

0.00

0.00

0.01

-0.0

10.

05-0

.02

0.05

-0.0

30.

05-0

.04

Rea

lexc

hang

era

te0.

140.

730.

150.

760.

160.

780.

220.

760.

250.

750.

260.

75

Num

ber

ofep

isod

es

Sove

reig

nde

faul

t0

028

492

830

923

Sudd

enst

ops1

3214

142

538

738

436

9

Avg

.sta

tist

ics

Publ

icsa

ving

s/G

DP2

-16.

1-1

8.9

-21.

2-2

0.8

-20.

6-2

0.5

Rea

lGD

Pgr

owth

33.

53.

53.

53.

63.

73.

7R

eale

xcha

nge

rate

101.

910

3.0

104.

310

7.8

110.

411

0.9

Publ

iccr

edit

/tra

dabl

eou

tput

40.

00.

00.

00.

00.

00.

0

Seri

esar

ede

tren

ded

usin

gH

Pfil

ter

wit

hpa

ram

eter

λ=

6.25

,so

stat

isti

csar

eca

lcul

ated

usin

gth

ede

viat

ions

from

the

tren

d.1

Con

side

rsth

ose

epis

odes

whe

reth

ebo

rrow

ing

limit

isbi

ndin

g.2

Con

side

rsth

epe

riod

sw

hen

ther

eis

full

acce

ssto

inte

rnat

iona

lcap

ital

mar

kets

.3

For

this

stat

isti

c,th

eG

DP

isde

flate

dby

the

cons

umer

pric

ein

dex

ofth

eec

onom

y.4

Con

side

rsth

ose

epis

odes

whe

reth

ebo

rrow

ing

limit

has

alo

wre

alis

atio

n.

48

Documentos de Trabajo

Banco Central de Chile

NÚMEROS ANTERIORES

La serie de Documentos de Trabajo en versión PDF

puede obtenerse gratis en la dirección electrónica:

www.bcentral.cl/esp/estpub/estudios/dtbc.

Existe la posibilidad de solicitar una copia impresa

con un costo de Ch$500 si es dentro de Chile y

US$12 si es fuera de Chile. Las solicitudes se

pueden hacer por fax: +56 2 26702231 o a través del

correo electrónico: bcch@bcentral.cl.

Working Papers

Central Bank of Chile

PAST ISSUES

Working Papers in PDF format can be downloaded

free of charge from:

www.bcentral.cl/eng/stdpub/studies/workingpaper.

Printed versions can be ordered individually for

US$12 per copy (for order inside Chile the charge

is Ch$500.) Orders can be placed by fax: +56 2

26702231 or by email: bcch@bcentral.cl.

DTBC – 815

Desarrollo del Crowdfunding en Chile

Iván Abarca

DTBC – 814

Expectativas Financieras y Tasas Forward en Chile

Rodrigo Alfaro, Antonio Fernandois y Andrés Sagner

DTBC – 813

Identifying Complex Core-Periphery Structures in the Interbank Market

José Gabriel Carreño y Rodrigo Cifuentes

DTBC – 812

Labor Market Flows: Evidence for Chile Using Micro Data from Administrative Tax

Records Elías Albagli, Alejandra Chovar, Emiliano Luttini, Carlos Madeira, Alberto Naudon,

Matías Tapia

DTBC – 811

An Overview of Inflation-Targeting Frameworks: Institutional Arrangements,

Decision-making, & the Communication of Monetary Policy Alberto Naudon y Andrés Pérez

DTBC – 810

How do manufacturing exports react to RER and foreign demand? The Chilean case

Jorge Fornero, Miguel Fuentes y Andrés Gatty

DTBC – 809

A Model of Labor Supply, Fixed Costs and Work Schedules

Gonzalo Castex y Evgenia Detcher

DTBC – 808

Dispersed Information and Sovereign Risk Premia

Paula Margaretic y Sebastián Becerra

DTBC – 807

The Implications of Exhaustible Resources and Sectoral Composition for Growth

Accounting: An Application to Chile

Claudia De La Huerta y Emiliano Luttini

DTBC – 806

Distribución de Riqueza No Previsional de los Hogares Chilenos

Felipe Martínez y Francisca Uribe

DTBC – 805

Geopolitical Tensions, OPEC News, and Oil Price: A Granger Causality Analysis

Carlos Medel

DTBC – 804

Spillovers and Relationships in Cross-Border Banking: The Case of Chile

Andrés Alegría, Kevin Cowan y Pablo García

DTBC – 803

Determinantes de la Inflación de Servicios en Chile

Mario Marcel, Carlos Medel y Jessica Mena

DTBC – 802

Banks’ Lending Growth in Chile: The Role of the Senior Loan Officers Survey

Alejandro Jara, Juan-Francisco Martínez y Daniel Oda

DTBC – 801

Pruebas de Tensión Bancaria del Banco Central de Chile: Actualización

Juan-Francisco Martínez, Rodrigo Cifuentes y J. Sebastián Becerra

DOCUMENTOS DE TRABAJO • Abril 2018